Frequent questions
AQL stands for Acceptable Quality Level
The concept stems from statistical analysis, and in particular the representativity of a population by a randomly drawn sample.
Take a random sample. Depending on its size it will exhibit the same characteristics as the general population, to a degree of accuracy (Think of political polling - the same formulas are used). The larger the random sample the more accurate its ability to predict the characteritics of the general population.
Back to Quality Control: AQL tables help define, for a requested degree of accuracy, the acceptable defect levels that can be found in the sample. Defects are usually divided into 3 categories (Critical, Major and Minor), and acceptance thresholds are defined for each.
Based on this simple information we will define rules for your product batch. Random sampling QC results will define whether the whole batch passes or fails your criteria. This decides if you should accept the shipment, ask the factory for a rework or discount, etc.
Please download the full AQL tables for a view of:
- sample quantities for lot sizes
- defect category statistical choices
- accept / reject criteria depending on defect category
Notes:
- The key to good sampling is RANDOM sampling.
- A slightly perverse effect of AQL is the assumption that perfection is impossible: seeking perfection is a waste of time: rejects are part of life and one just makes sure they remain manageable. To illustrate the effect, think of the Western car industry in the 1970's vs Japanese competition. This is why Quality Management in factory management culture makes such a huge difference. AQL is the way for buyers to check a one-off order, but if you have an ongoing relationship with a supplier then Quality Assurance, TQM, root cause analysis, continuous improvement feedback loops and Six Sigma are the way forward.
- Please contact us to define the method best suited to your needs.
